How does the client identify itself during the SMTP or IMAP transaction?

It uses the password that you generated for it. I don't understand where the confusion is.

I don't understand how that's any more secure than just using a strong password for the account. At some point, you're going to have to make that password accessible to the client. Plus, it's arguably less secure because the account now has multiple valid passwords that will work for authentication, and, based on your description, there's nothing that prevents someone from using the exact same password over a netcat session from accessing the account.

The confusion seems to be about logging into your account on the web versus using a mail client like Outlook or Thunderbird.

Pick a service that lets you use a long password and a security key (like Yubikey) or authenticator (Google, Authy) to log in.

Most services will then let you generate a specific password for an email client. I would assume that behind the scenes that the service is restricting what ports that password can be used on, etc.

> I would assume that behind the scenes that the service is restricting what ports that password can be used on

Assuming it's a device accessing the service over IMAP and SMTP that can access multiple networks, restricting by IP and/or port won't really help. As I noted in my other reply, it's easy enough to script access to the account if have the password and there's no real association between the application and the credentials that are used for access.